Fire detection using stereoscopic imaging and image processing techniques

This paper presents a stereoscopic imaging technique for fire detection. A low-cost binocular CCD camera is used to acquire the stereoscopic images of a controlled gas fire. Computer algorithms are developed to segment the fire zone in the HSV (Hue, Saturation, Value) color space and to determine the correspondence points of image pairs. The location, size and propagation of the fire are then determined from the perspective relationship between the fire and the camera. Experimental results under laboratory conditions show the effectiveness of the proposed approach for fire detection.